Perspectives on salvage treatment in prostate cancer
Salvage treatment represents the second radical treatment after the first attempt has failed. Up to 38% of patients will be informed after primary radical prostatectomy (P-RP) that their margins were positive, that recurrence is likely and that adjuvant salvage radiotherapy (S-RT) should be considered . The percentage of patients with evidence of recurrence after primary radiotherapy (P-RT) depends on whether and how follow-up biopsies are taken. Positive biopsy rates as high as 67% have been reported . Ablative technologies should not be considered as salvage therapy methods unless they are recognized as a true primary treatment. High-intensity focused ultrasonography and cryoablation may be used as primary treatments in carefully selected patients, although only cryoablation is recommended in the current European Association of Urology guidelines .
While there is high level evidence for the oncological effect of P-RP, the use of P-RT is based on consensus between experts and the evidence for oncological effect is poor . The evidence for salvage treatment is the opposite, with high level evidence for the effectiveness of S-RT where there are positive margins after RP, and poor evidence for the oncological effect of salvage treatment after failed RT. Interestingly, urologists question the clinical benefit of S-RT  and radiologists are sceptical about S-RP .
Weighing benefits and risks of salvage treatment is difficult, not only because the evidence for an oncological benefit is poor, but also because side effects may be serious; however, when it comes to S-RP and salvage cryoablation after P-RT, numerous patient series have shown results similar to those of primary treatment [7, 8]. Cryoablation seems to have the most favourable outcomes in terms of side effects. Far too often, however, salvage treatment is not considered and a rising PSA level is managed with androgen deprivation therapy (ADT) . The side effects of ADT must not be ignored, among them a 70% increased risk of serious cardiovascular events . Deferring ADT has become a valid aim of salvage treatment.
A successful outcome of primary as well as salvage treatment depends on giving the right treatment to the right patient by the right surgeon [7-9]. An expert surgeon not only perfects his operating technique, but is also allied to appropriate patient selection. The most important predictors for a successful outcome of salvage treatment are Gleason score, pretreatment PSA level, T-stage and PSA doubling time [7-9]. The importance of Gleason score and PSA is shown by Spiess et al. . While other predictors are static, PSA doubling time is dynamic and indicates how fast the tumour is growing, which is really what matters for primary as well as salvage treatment. Spiess et al. also show that PSA nadir after treatment provides guidance on which patients will need future ADT.
The mutual scepticism of radiologists and urologists to salvage treatment by the other party is justified and reflects our limited understanding of prostate cancer growth kinetics and their clinical consequences. Indeed, tumour growth rate and the time when spread occurs are the common denominators of all selection criteria. With a constant cell doubling time the amount of tumour tissue will rise exponentially from a certain time point, as will PSA doubling time (Fig. 1). That is why time is more important for salvage treatment than it is before primary treatment.
Measures to improve our ability to prognosticate and reduce the number of primary treatment failures are important for improving our care of patients with prostate cancer. Pretreatment assessment should focus on the tumour growth rate and the extent of tumour tissue in the body. More biopsies should be taken after P-RT to enable us to give salvage treatment without delay and to learn the true outcome of P-RT. ADT should not be given until salvage treatment has been considered.
Spiess et al. are to be congratulated on the achievements reached through the COLD registry [9, 10]. A registry is an important means of obtaining evidence in support of new treatment methods for prostate cancer. Comparing new methods with RP in randomized controlled trials might require large numbers of patients and >10 years observation time. This means most new methods are never likely to be recognized as an effective treatment option, whether primary or salvage. If a new non-extirpative method can demonstrate cancer cell death in biopsies and side effects are acceptable, then long-term controlled studies are less important. Future clinical studies as well as our practice of diagnosis and treatment must go hand in hand with a search for a better understanding of the dynamics of prostate cancer growth.